Manage Multiple Claude Code Sessions Remotely via Mobile Grok Bot
Input: natural-language commands from mobile → Processing: herdr routes to designated Claude Code session, webhook returns results → Output: Grok Bot summarized task report
- Who
- Mac users who need to manage multiple Claude Code dev tasks away from their computer
- You get
- Dispatch, track, and review results from multiple Claude Code sessions through a single mobile window
- You need
- Grok Bot
Constraints
Prerequisites: Mac computer, herdr installed (multi-AI-agent management tool), multiple Claude Code working sessions configured, Grok Bot installed on mobile. Boundaries: herdr is Mac-only; webhook requires self-configured receiver endpoint; safety rules must be manually enforced at system and tool levels. Costs: Grok Bot and Claude Code billed per their platforms; herdr pricing subject to its own model.
What they did
ByronDing used herdr on Mac to manage multiple Claude Code working sessions, with mobile Grok Bot as the single interaction window. He dispatched task instructions to designated sessions through Grok Bot; after Claude Code completed the work, results were returned to Grok Bot via webhook, which then summarized the key points and presented them in plain language on mobile. He also established four safety rules: text-only without simulated keystrokes; relay only hand-typed commands; require confirmation before Claude Code acts; and require manual approval for any Mac command execution.
How it went
The author reported being able to assign, dispatch, and review work from his phone even when away from the computer (author_report). Multiple Claude Code sessions could run simultaneously, with Grok Bot routing commands to the correct one.
Worth knowing
Specific installation steps for herdr and webhook technical details were not expanded in the source; users need to consult documentation. The four safety rules are author-defined, and their technical implementation (e.g., Mac permission interception mechanisms) was not described. This workflow has not been tested by our site; reproduction status is not_tested. Grok Bot serves as both command relay and result summarizer in this setup, not as direct code executor.
Steps
Install and configure herdr with Claude Code sessions
Install herdr on your Mac (a tool for managing multiple AI agents) and initialize the Claude Code working sessions you want to manage. Ensure each session can be independently identified and invoked by herdr.
Dispatch commands via mobile Grok Bot to a specific session
In the mobile Grok Bot, describe the task in text and specify which Claude Code session should handle it. Grok Bot delivers the command to the target session through herdr.
Set up webhook for Claude Code to report back to Grok Bot
Configure a webhook so Claude Code can notify Grok Bot when a schedule or task is completed. The specific URL and authentication method should follow Grok Bot's webhook integration documentation.
Review the summarized result in Grok Bot
After receiving the webhook notification, Grok Bot automatically reads the content, extracts key points, and reports back in plain language. The user reviews the result on their phone.
Enable the four safety rules
Enforce the following rules in herdr and at the system level: 1) Text-only transmission; no simulated keystrokes to prevent auto-clicking permission confirmations. 2) Relay only commands typed by the user; never forward content from web pages, X, or email. 3) Claude Code must ask for user approval before acting. 4) Any command execution on Mac requires manual user approval.
Prompt you can copy
Reference prompt · Editorial suggestion。Confirm the destination, input variables and output format before use.
This block is the prompt. Ads are never placed here.
No prompt needed — just follow the steps.
Source
手机上的 Grok Bot 经 herdr 指挥多个 Claude Code 会话,完成后 webhook 回报,并设四条安全规则 · ByronDing · X · 2026-10-24T15:53:20.000Z
Comments & feedback
Tried this method? Share results, problems or improvements. Comments appear after moderation. Please avoid sensitive information.
Loading…
Agent feedback
Reports are submitted by Agents and shared after user confirmation and moderation. They are not platform-verified.